.
T

rade and technology development policies almost always have distributional consequences. There may be a few exceptions for which the implementation of a policy produces either gains or no loss for nearly everyone, what economists would call a Pareto improvement. But these instances are relatively rare. You could argue that for early-stage developing countries, the export-driven growth model that draws surplus labor into the modernizing manufacturing and urban sectors comes close to meeting this standard. But even there, the gains are not spread evenly, and income inequality normally increases.

Distributional impacts are the norm, within countries and across national boundaries. Successful developing countries experience structural change as part of the growth process. The long-term benefits of exposure to global markets and investment are very large, driving both growth and significant structural adjustments in terms of jobs, skills, and human capital. But some sectors are inevitably adversely affected.

To ensure that new economic opportunities and pressures do not overwhelm the ability of developing countries–particularly the labor force–to adapt, policymakers should manage the pace and sequencing of the opening process in trade, investment, and the capital account. For example, if net employment creation–jobs created minus jobs lost–turns negative, opening may be happening too fast.

Efforts to calibrate the pace of opening should be complemented by some redistribution toward adversely affected people or sectors, but not at the expense of investment. Most important, to support the creation of an inclusive pattern of structural adjustment, government must invest heavily in high-quality, affordable (either low-cost or free) education for young people and training for older workers. 

All of this is vital to ensure that the policies that underpin the growth model retain popular support; otherwise, political opposition will likely disrupt or even abort the growth strategy.

These challenges are not limited to developing economies. Trade, investment, and technology have significant effects on economic structure, relative prices, and income and wealth distribution pretty much everywhere. One recent paper argues that trade with China not only has direct negative effects on employment and wages in the US manufacturing sector, but also produces negative upstream effects on suppliers of intermediate products.

To be sure, the paper’s authors conclude that, for the United States, trade with China yields net benefits, because the positive downstream effect–a wide range of industries gaining access to cheaper intermediate products – is larger than the combined direct and upstream negative effects. Nevertheless, U.S.-China trade still has important distributional implications because the negative effects are more concentrated by sector and geography, whereas the positive effects are spread widely. This has arguably had a significant impact on American attitudes toward trade with China-and thus on U.S. trade policy generally.

Of course, the debate about trade with China is particularly heated in the U.S., owing not least to allegations that China has violated World Trade Organization rules. But this is a red herring. There are surely many cases of developing countries’ failure to comply strictly with WTO rules. But the structural and distributional effects of trade do not depend on a country’s compliance with WTO rules, but rather on its stage of development, the scale of trade, and its comparative advantages.

The severity of the so-called China shock in the U.S. reflected its speed and scale. Policymakers’ mistake was to devote relatively little attention to modulating the speed of the transition or supporting those affected by the structural adjustment.

But slowing the pace of structural change is easier said than done, particularly when it comes to the green transition–another key driver of structural change today. Decades of inaction mean that rapid reductions in greenhouse-gas emissions are now urgently needed. But this is already creating major dislocations, with serious distributional implications. As these effects grow, so will resistance to the necessary initiatives.

A third driver of structural transformation today is technology. As David Autor and others have documented, even before the latest breakthroughs in artificial intelligence, digital technology was removing routine (codifiable), mainly middle-income jobs from the economy, leading to job and income polarization. This phenomenon can be observed in all advanced economies.

Compounding the challenge in the US, productivity growth has moved onto a dual track. As Belinda Azenui and I recently noted, breakthroughs in machine learning have enabled productivity to grow rapidly in what technologists call the “bits layer” of the economy–where information is processed, stored, accessed, and used, where transactions occur, and where decisions are made.

But in the “atoms layer,” where physical economic activity takes place, productivity growth is mixed–higher in structured environments like manufacturing and logistics, and lower elsewhere, including large employment sectors like hospitality. If these trends–and policymakers’ inaction–persist, the gap in productivity and incomes will continue to widen.

In a 2022 article entitled “The Turing Trap,” Erik Brynjolfsson suggested that the AI research agenda is overly focused on human-like artificial intelligence, motivated by the famous Turing Test: can a person interacting with a machine determine whether it is one? Clearly, that benchmark has produced astonishing advances. But Brynjolfsson argues that it needs to be complemented with a more aggressive and well-funded machine-augmentation agenda. The goal of developing semi-autonomous vehicles must be accompanied by a push to boost the productivity of a broad range of service-sector jobs. 

Copyright: Project Syndicate, 2023.

About
Michael Spence
:
Michael Spence, a Nobel laureate in economics, is Emeritus Professor of Economics and a former dean of the Graduate School of Business at Stanford University.
The views presented in this article are the author’s own and do not necessarily represent the views of any other organization.

a global affairs media network

www.diplomaticourier.com

Countering Structural Disruptions

Image by Bellergy RC from Pixabay

February 23, 2023

Trade and technology development policies usually lead to distributional impacts, causing distruption or unintended harms to segments of the population. Our fascination with developing human-like AI at the cost of a machine augmentation agenda makes this worse, writes Nobel Laureate Michael Spence.

T

rade and technology development policies almost always have distributional consequences. There may be a few exceptions for which the implementation of a policy produces either gains or no loss for nearly everyone, what economists would call a Pareto improvement. But these instances are relatively rare. You could argue that for early-stage developing countries, the export-driven growth model that draws surplus labor into the modernizing manufacturing and urban sectors comes close to meeting this standard. But even there, the gains are not spread evenly, and income inequality normally increases.

Distributional impacts are the norm, within countries and across national boundaries. Successful developing countries experience structural change as part of the growth process. The long-term benefits of exposure to global markets and investment are very large, driving both growth and significant structural adjustments in terms of jobs, skills, and human capital. But some sectors are inevitably adversely affected.

To ensure that new economic opportunities and pressures do not overwhelm the ability of developing countries–particularly the labor force–to adapt, policymakers should manage the pace and sequencing of the opening process in trade, investment, and the capital account. For example, if net employment creation–jobs created minus jobs lost–turns negative, opening may be happening too fast.

Efforts to calibrate the pace of opening should be complemented by some redistribution toward adversely affected people or sectors, but not at the expense of investment. Most important, to support the creation of an inclusive pattern of structural adjustment, government must invest heavily in high-quality, affordable (either low-cost or free) education for young people and training for older workers. 

All of this is vital to ensure that the policies that underpin the growth model retain popular support; otherwise, political opposition will likely disrupt or even abort the growth strategy.

These challenges are not limited to developing economies. Trade, investment, and technology have significant effects on economic structure, relative prices, and income and wealth distribution pretty much everywhere. One recent paper argues that trade with China not only has direct negative effects on employment and wages in the US manufacturing sector, but also produces negative upstream effects on suppliers of intermediate products.

To be sure, the paper’s authors conclude that, for the United States, trade with China yields net benefits, because the positive downstream effect–a wide range of industries gaining access to cheaper intermediate products – is larger than the combined direct and upstream negative effects. Nevertheless, U.S.-China trade still has important distributional implications because the negative effects are more concentrated by sector and geography, whereas the positive effects are spread widely. This has arguably had a significant impact on American attitudes toward trade with China-and thus on U.S. trade policy generally.

Of course, the debate about trade with China is particularly heated in the U.S., owing not least to allegations that China has violated World Trade Organization rules. But this is a red herring. There are surely many cases of developing countries’ failure to comply strictly with WTO rules. But the structural and distributional effects of trade do not depend on a country’s compliance with WTO rules, but rather on its stage of development, the scale of trade, and its comparative advantages.

The severity of the so-called China shock in the U.S. reflected its speed and scale. Policymakers’ mistake was to devote relatively little attention to modulating the speed of the transition or supporting those affected by the structural adjustment.

But slowing the pace of structural change is easier said than done, particularly when it comes to the green transition–another key driver of structural change today. Decades of inaction mean that rapid reductions in greenhouse-gas emissions are now urgently needed. But this is already creating major dislocations, with serious distributional implications. As these effects grow, so will resistance to the necessary initiatives.

A third driver of structural transformation today is technology. As David Autor and others have documented, even before the latest breakthroughs in artificial intelligence, digital technology was removing routine (codifiable), mainly middle-income jobs from the economy, leading to job and income polarization. This phenomenon can be observed in all advanced economies.

Compounding the challenge in the US, productivity growth has moved onto a dual track. As Belinda Azenui and I recently noted, breakthroughs in machine learning have enabled productivity to grow rapidly in what technologists call the “bits layer” of the economy–where information is processed, stored, accessed, and used, where transactions occur, and where decisions are made.

But in the “atoms layer,” where physical economic activity takes place, productivity growth is mixed–higher in structured environments like manufacturing and logistics, and lower elsewhere, including large employment sectors like hospitality. If these trends–and policymakers’ inaction–persist, the gap in productivity and incomes will continue to widen.

In a 2022 article entitled “The Turing Trap,” Erik Brynjolfsson suggested that the AI research agenda is overly focused on human-like artificial intelligence, motivated by the famous Turing Test: can a person interacting with a machine determine whether it is one? Clearly, that benchmark has produced astonishing advances. But Brynjolfsson argues that it needs to be complemented with a more aggressive and well-funded machine-augmentation agenda. The goal of developing semi-autonomous vehicles must be accompanied by a push to boost the productivity of a broad range of service-sector jobs. 

Copyright: Project Syndicate, 2023.

About
Michael Spence
:
Michael Spence, a Nobel laureate in economics, is Emeritus Professor of Economics and a former dean of the Graduate School of Business at Stanford University.
The views presented in this article are the author’s own and do not necessarily represent the views of any other organization.